The purification by affinity chromatography of a glutathione S-transferase from larvae of Galleria mellonella

Production and purification of homogenous recombinant human selenoproteins reveals a unique codon skipping event in E. coli and GPX4-specific affinity to bromosulfophthalein

Selenoproteins are translated via animal domain-specific elongation machineries that redefine dedicated UGA opal codons from termination of translation to selenocysteine (Sec) insertion, utilizing specific tRNA species and Sec-specific elongation factors. This has made recombinant production of mammalian selenoproteins in E. coli technically challenging but recently we developed a methodology that enables such production, using recoding of UAG for Sec in an RF1-deficient host strain. Here we used that approach for production of the human glutathione peroxidases 1, 2 and 4 (GPX1, GPX2 and GPX4), with all these three enzymes being important antioxidant selenoproteins. Among these, GPX4 is the sole embryonically essential enzyme, and is also known to be essential for spermatogenesis as well as protection from cell death through ferroptosis. Enzyme kinetics, ICP-MS and mass spectrometry analyses of the purified recombinant proteins were used to characterize selenoprotein characteristics and their Sec contents. This revealed a unique phenomenon of one-codon skipping, resulting in a lack of a single amino acid at the position corresponding to the selenocysteine (Sec) residue, in about 30% of the recombinant GPX isoenzyme products. We furthermore confirmed the previously described UAG suppression with Lys or Gln as well as a minor suppression with Tyr, together resulting in about 20% Sec contents in the full-length proteins. No additional frameshifts or translational errors were detected. We subsequently found that Sec-containing GPX4 could be further purified over a bromosulfophthalein-column, yielding purified recombinant GPX4 with close to complete Sec contents. This production method for homogenously purified GPX4 should help to further advance the studies of this important selenoprotein.

Manipulation of host ecdysteroid hormone levels facilitates infection by the fungal insect pathogen, Metarhizium rileyi

Entomopathogenic fungi such as Metarhizium rileyi and Beauveria bassiana are widely used insect biological control agents. Little, however, is known concerning genetic or enzymatic factors that differentiate the mechanisms employed by these two fungal pathogens to infect target hosts. Infection by either of these organisms is known to increase levels of the growth and molting hormone, ecdysone, which also regulates the expression of a number of innate immune pathways. M. rileyi, but not B. bassiana, has apparently evolved an ecdysteroid-22-oxidase (MrE22O) that inactivate ecdysone. We show that deletion of MrE22O impaired virulence compared with the wild-type strain, with an increase in ecdysone titer seen in hosts that was coupled to an increase in the expression of antimicrobial genes. An M. rileyi strain engineered to overexpress MrE22O (MrE22O^OE ), as well as trans-expression in B. bassiana (Bb::MrE220^OE ) resulted, in strains displaying enhanced virulence and dampening of host immune responses compared with their respective wild-type parental strains. These results indicate that ecdysone plays an important role in mediating responses to fungal infection and that some insect pathogenic fungi have evolved mechanisms for targeting this hormone as a means for facilitating infection.

The Dac-tag, an affinity tag based on penicillin-binding protein 5

Penicillin-binding protein 5 (PBP5), a product of the Escherichia coli gene dacA, possesses some β-lactamase activity. On binding to penicillin or related antibiotics via an ester bond, it deacylates and destroys them functionally by opening the β-lactam ring. This process takes several minutes. We exploited this process and showed that a fragment of PBP5 can be used as a reversible and monomeric affinity tag. At ambient temperature (e.g., 22°C), a PBP5 fragment binds rapidly and specifically to ampicillin Sepharose. Release can be facilitated either by eluting with 10mM ampicillin or in a ligand-free manner by incubation in the cold (1-10°C) in the presence of 5% glycerol. The "Dac-tag", named with reference to the gene dacA, allows the isolation of remarkably pure fusion protein from a wide variety of expression systems, including (in particular) eukaryotic expression systems.

Studies on the glutathione S-transferase proteome of adultDrosophila melanogaster: Responsiveness to chemical challenge

GSTs from adult Drosophila melanogaster have been partially purified using three different affinity chromatography media and separated by 2-DE. Nine GSTs have been identified by MALDI-TOF MS. In the absence of special treatments, eight GSTs could be positively identified. These were DmGSTs D1 (the dominant Delta isoform which was present in five protein zones of differing pI) and D3 (and possibly also D5); the Epsilon-class GSTs E3, 6, 7 and 9 and a previously uncharacterised, probable member of the class, CG16936. The Sigma-class DmGSTS1 was prominent. DmGSTD2 was detected only after pretreatment of the flies with Phenobarbital (PhB). Treatment with Paraquat (PQ) led to an increase in the total GST activity, as measured with the substrates 1-chloro-2,4-dinitrobenzene (CDNB) and 3,4-dichloro-nitrobenzene (DCNB) and an increase in the relative amounts of the D1, D3, E6 and E7 isoforms. PhB treatment led to increases in the relative amounts of the D1, D2, E3, E6, E7 and E9 isoforms detected with a possible depression in the relative amount of GSTS1. CG16936 was unaffected by either pretreatment.

Design of a monomeric human glutathione transferase GSTP1, a structurally stable but catalytically inactive protein

By the introduction of 10 site-specific mutations in the dimer interface of human glutathione transferase P1-1 (GSTP1-1), a stable monomeric protein variant, GSTP1, was obtained. The monomer had lost the catalytic activity but retained the affinity for a number of electrophilic compounds normally serving as substrates for GSTP1-1. Fluorescence and circular dichroism spectra of the monomer and wild-type proteins were similar, indicating that there are no large structural differences between the subunits of the respective proteins. The GSTs have potential as targets for in vitro evolution and redesign with the aim of developing proteins with novel properties. To this end, a monomeric GST variant may have distinct advantages.

Chromatographic separation of fluorescent thiol adducts of 4-chloro-7-sulphobenzofurazan. Use as substrates for enzymes of the mercapturic acid xenobiotic pathway

Fluorescent adducts of 4-chloro-7-sulphobenzofurazan with cysteine, cysteinylglycine, reduced glutathione and N-acetylcysteine were prepared. Adducts were separated by HPLC on a 3-mm Nova-Pak C18 reversed-phase column using isocratic elution with a solvent of acetonitrile-0.15 M phosphoric acid (5:95) buffered at pH 2.5. The adducts were detected using a fluorescence detector set at an excitation wavelength of 365 nm and an emission wavelength of 510 nm and an ultraviolet detector at 254 nm. The adduct of reduced glutathione was also formed by the action of the enzyme glutathione-S-transferase. This adduct acted as a substrate for the enzyme gamma-glutamyltranspeptidase and the product of this reaction, the 4-chloro-7-sulphobenzofurazanyl derivative of cysteinylglycine, acted as a substrate for either dipeptidase or aminopeptidase M. The sequential enzymic effects could be detected by changes in the relative fluorescence intensity of the solutions to which the respective enzymes had been added but were more appropriately followed by changes in the HPLC elution profiles after enzymic treatment of solutions.

Inhibition of glutathione S-transferases from rat liver by basic triphenylmethane dyes

(1) Basic triphenylmethane dyes related to pararosanilin inhibit class alpha glutathione S-transferases (GSTs) from rat liver. The inhibitory potency of each dye correlates with its octanol-water partition coefficient. Values of Ki determined at pH 6.5 ranged from about 1 x 10(-7) M for Ethyl violet to 7 x 10(-5) M for Methyl green. GST 3-3, a class mu isoenzyme, was an order of magnitude less sensitive to inhibition by Ethyl violet. (2) All of the dyes tested were bleached to varying degrees by glutathione. The bleaching appears to result from the formation of an adduct between the dye and glutathione. At pH 6.5, adduct formation is significant only for Malachite green and Methyl green. There is kinetic evidence that for these dyes the adduct contributes significantly to the overall inhibition. It is probable that at physiological pH, all of the dyes would exist to a significant extent in the adduct form. (3) The dyes are excreted extensively in the bile, at least partly as the glutathione adduct. The free dye is regenerated on standing, it is assumed as a result of removal of glutathione by oxidation.

Purified glutathione S-transferases from parasites as candidate protective antigens

Glutathione S-transferase (GST) activity was detected in larvae of the Australian sheep blowfly Lucilia cuprina, and in the nematode Haemonchus contortus. A specific inhibitor of the enzyme was shown to affect survival of both species of parasite in vitro. GST from both parasites has been purified and partially characterized. Antisera raised to the purified enzymes were shown to inhibit the enzyme activity in vitro. However, the antisera had no effect on the survival of either parasite.

Inhibition by inorganic anions of glutathione S-transferases from insect and mammalian sources

Glutathione S-transferases 1-1, 3-3, 3-4 and 4-4 from rat liver and the major glutathione S-transferase from the wax moth (Galleria mellonella) are all inhibited by several simple inorganic anions. For each of 3-3, 3-4 and the insect enzyme, the order of inhibitory potency was ClO4- greater than or equal to SCN- greater than I- greater than NO3- greater than Br-. A more limited range of anions was tested on the isoenzymes 1-1 and 4-4, but the same trend was apparent. Values for Ki ranged from about 200 mM for Cl- to 6 mM for SCN- in the case of the insect enzyme and from 50 mM for Br- to 0.3 mM for SCN- for the rat isoform 3-3. Acetate, F-, SO4(2-) and PO4(3-) were not found to have significant inhibitory properties. The mode of inhibition was characterized as non-competitive in the case of the insect enzyme and rat transferase 1-1, whereas the mode of inhibition was partially non-competitive in the case of the rat isoforms 3-3, 3-4 and 4-4.

Partial purification and isoelectric focusing patterns of the buffalo (Bubalus bubalis) and the Kedah-Kelantan cattle (Bos indicus) glutathione transferases

1. Glutathione transferases from the liver, lung and kidney tissues of the buffalo (Bubalus bubalis) and the Kedah-Kelantan cattle (Bos indicus) were partially purified by ammonium sulphate precipitation and Sephadex G-75 gel filtration. 2. Liver tissue contains the highest enzyme activity when compared to the lung and kidney tissues. 3. The activity in cattle is higher than that in the buffalo. 4. Isoelectric focusing separates the activities into the acidic, near neutral and basic fractions. 5. The focused patterns are different for each of the tissues and in each of the species investigated.

Synthesis and use of an isoform-specific affinity matrix in the purification of glutathione S-transferases from the housefly, Musca domestica (L.)

Glutathione may be linked to an agarose matrix which has been activated by treatment with epichlorhydrin. The resulting resin displayed group selectivity for the glutathione S-transferases of the housefly Musca domestica (L). The isoenzymes of low isoelectric point, which have little activity with substrates other than 1-chloro-2,4-dinitrobenzene, bound strongly to this matrix and were eluted with 10 mM glutathione at pH 7.4. On the other hand, the group of isoenzymes of higher isoelectric point, showing activity with other substrates such as 3,4-dichloronitrobenzene, did not bind. These isoenzymes did bind to a sulfobromophthalein-glutathione conjugate immobilized on agarose and could be eluted with 5 mM sulfobromophthalein at pH 7.4. The immobilized glutathione resin bound rat liver glutathione S-transferase subunits from all three molecular weight classes.

The Meisenheimer complex of glutathione and trinitrobenzene. A potent inhibitor of the glutathione S-transferase from Galleria mellonella

1. The Meisenheimer complex formed between reduced glutathione and 1,3,5-trinitrobenzene is characterised by an extinction coefficient at 470 nm of 20400 and by an association constant at pH 9.18 of 42 l.mol-1. 2. Trinitrobenzene is a moderately good inhibitor of the glutathione S-transferase from larvae of the moth Galleria mellonella. It acts by competition with the electrophilic substrate. At pH 7.4, it has a Ki value of 10 microM. Its mode of inhibition with respect to GSH appears to be non-competitive. 3. At pH values below 9.0, the Meisenheimer complex does not appear to be formed in sufficient quantity to give significant inhibition of the enzyme. At pH 9.0 and at GSH concentrations greater than 1 mM, the inhibition of the enzyme became markedly non-hyperbolic. This was attributed to the inhibitory action of the Meisenheimer complex. The complex appears to act also by competition with the electrophilic substrate and its Ki is calculated to be 1.7 X 10(-7) M.

The purification and characterization of glutathione S-transferase from the hepatopancreas of the blue crab, Callinectes sapidus

High glutathione S-transferase activity was found in the cytosol of F-cells from the hepatopancreas of the blue crab (Callinectes sapidus). Purification of glutathione S-transferase from hepatopancreas extracts by Sephadex G-200, DEAE-Sephacel, and chromatofocusing resulted in the isolation of two isozymes with isoelectric points of 5.9 and 5.7, as determined by analytical isoelectric focusing. Using 1-chloro-2,4-dinitrobenzene as the substrate the specific activities of the two purified isozymes were 222 and 182 mumol/min/mg, respectively. There was no evidence for basic transferase isozymes. In addition to 1-chloro-2,4-dinitrobenzene the purified glutathione S-transferase isozymes showed activity with p-nitrophenyl acetate, p-nitrobenzyl chloride, bromosulfophthalein, and benzopyrene oxide. Thus, both substitution and addition reactions associated with vertebrate glutathione S-transferase were found in the crab transferases. There was no when ethacrynic acid, methyl iodide, trans-4-phenyl-3-buten-2-one, 1,2-epoxy-(p-nitrophenoxy)propane, cumene hydroperoxide, and t-butyl hydroperoxide were used as substrates. The lack of peroxidase activity is of interest since this activity is commonly found in vertebrate transferase isozymes. The two transferases had a dimeric Mr of 40,800 with similar amino acid compositions and similar kinetic parameters (Vmax, Km, and pH maxima) with 1-chloro-2,4-dinitrobenzene as substrate. The two transferases could be distinguished by their isoelectric points, molecular mass of the monomers (22,300 for GST 1 and 22,300 and 22,400 for GST 2), and different inhibitor mechanisms with hematin and bromosulfophthalein.

Human erythrocyte glutathione S-transferase: a possible marker of chemical exposure

We have explored the possibility of using glutathione S-transferase (GST) as a biological marker of chemical exposure. All the model compounds tested in the present study (acrolein, propylene oxide, styrene oxide, ethylene dibromide and ethylene dichloride) showed a dose-dependent inactivation of erythrocyte GST in situ as well as the inhibition of purified erythrocyte GST.

Identification of multiple glutathione S-transferases from Daphnia magna

1. Six anionic glutathione S-transferases (GST) were purified from the crustacean, Daphnia magna, by means of affinity chromatography, that are responsible for ca. 40% of cytosolic GST activity. 2. Electrophoresis in the presence of sodium dodecyl sulfate (SDS) revealed the presence of three proteins, with molecular weights of 27,500, 28,000, and 30,200. 3. Separation under nondenaturing conditions revealed six proteins, all of which exhibited GST activity, with molecular weights ranging from 55,000 to 61,700. 4. Ethacrynic acid is a competitive inhibitor of activity towards CDNB of all six GSTs, binding each with similar affinities. 5. Chlorinated phenols are also competitive inhibitors of the enzyme, with the degree of inhibition being directly correlated with the lipophilicity of the compounds. 6. Analysis of inhibition of separated isoforms reveals that form 4 is most strongly inhibited by these chlorinated phenols, with forms 5 and 6 being inhibited to a lesser degree.

Anion-exchange high-performance liquid chromatography of glutathione S-transferases. Separation of the minor isoenzymes of human erythrocyte, heart and lung

Apparently homogeneous preparation of the acidic glutathione S-transferases were subjected to high-performance liquid chromatography over a SynChropak AX-300 anion-exchange column. Results of high-performance liquid chromatography analysis indicated that in each of these tissues the acidic glutathione S-transferases have one major and two minor isoenzymes. All the isoenzymes obtained by high-performance liquid chromatography are dimers of subunits of molecular weight (Mr) 22,500. The major isoenzymes of all the tissues, representing more than 80% of total glutathione S-transferase activity, have similar retention times and comparable kinetic properties. High-performance liquid chromatography profiles of the lung and heart isoenzymes are similar to each other but are significantly different from that of the erythrocytes. These studies provide evidence of microheterogeneity in the acidic isoenzymes of heart, lung and erythrocyte glutathione S-transferases and provide a rapid and efficient method for separation of minor isoenzymes in more than 95% yield.

Suppression of high-affinity ligand binding to the major glutathione S-transferase from Galleria mellonella by physiological concentrations of glutathione

The major glutathione S-transferase from larvae of Galleria mellonella binds a number of synthetic triphenylmethane dyes with dissociation constants of the order of 10(-6) M or less. The organ distribution of the enzyme activity does not parallel the uptake of such dyes by the insect's organs in vivo. The affinity of the protein for such dyes is decreased by about an order of magnitude by the presence of glutathione in normal physiological concentration. This appears to be the cause of this protein's lack of efficacy as a 'ligandin' in vivo. The dyes appear to be acting as ineffective substrate analogues, binding at the catalytic site and impeding, in a reciprocal fashion, the binding of glutathione. Fluorescence-quenching titration and kinetic experiments together indicate the existence of a single ligand-binding and catalytic site per dimeric enzyme molecule.

Distribution and properties of glutathione S-transferase from T. infestans

The glutathione transferase from T. infestans is able to render aqueous metabolites when incubated in vitro with malathion, parathion and fenitrothion. It is a soluble enzyme present in every developmental stage and widely distributed in all insect organs. The purification procedure applied, consisting of fractionation with ammonium sulfate and Bio-Gel P-60 chromatography, gives an unique molecular form catalytically active using methyl iodide as substrate in polyacrylamide gel electrophoresis (PAGE). One of the most active substrates is the 1-chloro-2,4-dinitrobenzene (CDNB), with an activity maximum at pH 7.5 and at 45 degrees C temperature. Its activation energy calculated from an Arrhenius plot is 14,846 cal mol-1. The enzyme susceptibility to inhibition by thiol reagents shows three degrees of responses; slight, moderate or high, depending on the compounds used. The kinetics of the enzyme catalysed reaction with the purified fraction is complex, and resembles that reported for glutathione S-transferase A from rat liver, showing a biphasic kinetic mechanism in which the reaction pathway depends on the concentration of GSH. In general, the properties of this insect enzyme are similar to those enzymes isolated from vertebrate organisms.

Purification and characterization of a new form of glutathione S-transferase from human erythrocytes

Presence of a new form of glutathione S-transferase has been demonstrated in human erythrocytes. Using two different affinity ligands this enzyme has been separated from the previously characterized glutathione S-transferases rho. The new enzyme is highly basic with a pI of greater than 10. The new enzyme which represents less than 5 percent of glutathione-S-transferase activity towards 1-chloro-2,4-dinitrobenzene as substrate and about 10 percent of total glutathione S-transferase protein of erythrocytes has different amino acid composition, substrate specificities, and immunological characteristics from those of the major erythrocyte glutathione S-transferase rho. Immunological properties of the new enzyme indicate that this form may be different from other glutathione S-transferases of human tissues.

The hepatic glutathione transferases of the male little skate, Raja erinacea

Five cytosolic glutathione transferases were isolated from the liver of the male little skate, Raja erinacea, a marine elasmobranch. They were designated E-1 through E-5 in order of their elution from a DEAE-cellulose column with a 0 to 100 mM KCl gradient in 0.01 M Tris (pH 8.0). Each eluted peak of glutathione transferase activity, after concentration, was applied to an affinity column prepared by reaction of epoxy-activated Sepharose 6B with glutathione (GSH). Elution of the various glutathione transferases from this column with GSH resulted in the further purification of each enzyme; the major glutathione transferase, E-4 and E-1, were purified to apparent homogeneity by this procedure. Skate glutathione transferase E-4 is dimeric and the subunits are either very similar or identical in molecular weight (about 26 000 daltons). Enzymes E-2 through E-5 were acidic proteins (pI less than 7.0) and had high specific glutathione transferase activity (0.3--12 mumol/min/mg protein) with benzo[a]pyrene 4,5-oxide (BPO) as substrate, whereas the other enzyme (E-1) had low activity (0.01 mumol/min/mg) with BPO and a basic pI (greater than 9.5). Bilirubin and hematin, non-substrate ligands, bound tightly to homogeneous E-4, with dissociation constants in the micromolar range.

Purification and properties of glutathione S-transferases from larvae of Wiseana cervinata

The glutathione S-transferases from the porina moth, Wiseanna cervinata, were purified by affinity chromatography, cation-exchange chromatography and preparative isoelectrofocusing. The major transferase (IV) was purified to homogeneity by a factor of 530-fold with a yield of 83%. Other transferases present were purified to a smaller degree (approx. 50-fold) to a stage of near-homogeneity. The transferases examined all had Mr values about 45 000-50 000. They appeared to be homodimers of either of two types of subunit, of Mr 22 800 and 24 600. Enzymes consisting of the different types of subunit were not immunologically cross-reactive. The major enzyme fractions separated by cation-exchange chromatography were both active with 1-chloro-2,4-dinitrobenzene, 1,2-dichloro-4-nitrobenzene, ethacrynic acid and iodomethane, but were inactive with 4-nitropyridine N-oxide, 1,2-epoxy-3-(p-nitrophenoxy)propane, bromosulphophthalein and p-nitrobenzyl chloride. The kinetics of the enzyme-catalysed reaction with enzyme IV were non-Michaelean with respect to both substrates. Both products were inhibitory. The results appear to be compatible with a random steady-state mechanism. It is concluded that these enzymes are very similar, in their physical and chemical constitution, in their catalytic properties and in their relationships with each other, to those enzymes that have been isolated from vertebrate organisms.

Purification from pig liver of a protein which protects liposomes and biomembranes from peroxidative degradation and exhibits glutathione peroxidase activity on phosphatidylcholine hydroperoxides

The cell sap from pig liver contains a protein which protects phosphatidylcholine liposomes and biomembranes from peroxidative degradation in the presence of glutathione. The activity of this protein has been assayed by measuring the inhibition of aged phosphatidylcholine liposome peroxidation induced by the Fe3+-triethylenetetramine complex. The peroxidation-inhibiting protein from pig liver has been purified 585-fold to homogeneity with overall recovery of activity of 12%. (NH4)2SO4 precipitation, ion-exchange chromatography on DEAE-Sepharose CL-6B and CM23-cellulose, affinity chromatography on glutathione-bromosulfophthalein-Sepharose and gel filtration on Sephadex G-50 were used. Gel filtration and SDS- polyacrylamide gel electrophoresis indicated a molecular weight of approximately 20 000. The protein inhibited peroxidation by Fe3+-triethylenetetramine following a 15 min preincubation of phosphatidylcholine liposomes in the presence of 5mM glutathione or 2-mercapthoethanol. The pure protein exhibited glutathione peroxidase activity on hydroperoxide groups of phosphatidylcholine and on cumene and t-butyl hydroperoxides, with specific activities of 2.2, 3.8 and 0.9 mumol/min per mg protein, respectively. The protein appears to be distinct from the selenoenzyme glutathione peroxidase and from any known glutathione S-transferase. The peroxidation was studied also with fresh phosphatidylcholine liposomes and was induced in this case by Fe-ascorbate. To obtain protection by the peroxidation-inhibiting protein and glutathione, preincubation was not necessary, but alpha-tocopherol, incorporated in the liposomes in the molar ratio 1:250 to phosphatidylcholine, was required. Lipid peroxidation of rat liver mitoplasts and microsomes was blocked when these preparations were incubated in the peroxidizing mixture in the presence of peroxidation-inhibiting protein and glutathione. The protection from Fe3+-triethylenetetramine-induced peroxidation is related apparently to reduction of hydroperoxide groups in polyunsaturated fatty acid residues of phospholipids and to inhibition of free radicals formation by chain branching. Protection from the Fe-ascorbate-induced peroxidation is apparently attributable to the same mechanism. However, the requirement of alpha-tocopherol for protection in the Fe-ascorbate-induced peroxidation suggests that the cooperation of a free-radical scavenger is necessary. It is probable that the glutathione peroxidase activity is involved also in the glutathione-dependent protection exhibited by the protein on lipid peroxidation of biomembranes.

Purification and characterization of an acid glutathione S-transferase from human lung

An acid glutathione S-transferase (EC: 2.5.1.18) from human lung was purified and characterized. The purification procedure included two isoelectric focusing runs, Sephadex G-100 gel filtration, glutathione-affinity chromatography, and Sephadex G-75 gel filtration. With respect to the properties studied the acid lung transferase differed from human liver transferases alpha-epsilon, but it bore a close resemblance to the other human low pI transferases. Bilirubin affected the kinetics of the lung enzyme markedly differently as compared with transferase p, suggesting possible nonidentity between these enzymes. The acid lung transferase represented about 97% of the total glutathione transferase activity of the lung 100,000 g supernatant used in this work.

Affinity chromatography of hepatic glutathione S-transferases on omega-aminoalkyl sepharose derivatives of glutathione

Rat liver glutathione S-transferases (RX: glutathione R-transferase, EC 2.5.1.18) were found to adsorb S-carbamidomethyl glutathione linked to Sepharose CL-4B via lysyl or aliphatic diamine spacers of various carbon chain lengths (-NH-(CH2)n-NH-, n = 2, 4, 5, 6, 8 and 10). Proteins were eluted specifically by reduced glutathione. The affinity of the enzymes for the adsorbent increased with increase in the carbon chain length of aliphatic diamine spacers used. Adsorbent having a free carboxyl group within the spacer moiety had high capacity and was specific for glutathione S-transferases. The transferases were specifically eluted from the column in high yield by low concentrations of glutathione. Enzymes purified by the lysyl spacer adsorbent were homogeneous in sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis and contained most of the hepatic glutathione S-transferase isozymes in isoelectric focusing. Oxidized glutathione and S-methyl glutathione were equally effective as reduced glutathione in eluting glutathione S-transferases from the adsorbent, but gamma-glutamylcysteinylglycineamide or gamma-glutamylcysteinylglycine-1-methyl ester were not effective. These data suggested that the free carboxyl group of glycyl moiety of glutathione might also be important for the specific binding of the transferases to this adsorbent.